Efficientnet for Brain Tumor Detection from MRI

The human brain is the primary controller of the humanoid system. The unusual expansion of the brain tissues leads to brain tumor. The continuous escalation of brain tissue leads to brain cancer. Computer vision plays an inevitable role in the field of medical science, and, in it, magnetic resonance imaging techniques are used to detect brain tumors. In the realm of image categorization, deep learning is a core topic. It currently has quite a promising potential in terms of brain tumor classification and segmentation. This work’s key principle is to build a deep convoultional neural network for detecting brain tumors. In the proposed model, the tumor region is first segmented from the MR images. Second, data augmentation is used to allow effective training, and, subsequently, a fine-tuned model EfficientNet is used for detecting multi-class brain tumor. The model is trained using brain tumor dataset. The method achieved an average accuracy of 97.35%.